: This study will utilize population-based molecular epidemiology to identify and evaluate tuberculosis control strategies that are appropriate under different epidemiologic circumstances. In addition, we will test several different methods for genotyping strains of Mycobacterium tuberculosis and will determine the optimum combination of techniques that provide the best balance of sensitivity and specificity for tracing the organism in a population. The specific aims of the project are grouped according to the component of the study to which they relate. Component 1: 1) To use 1S6110-based RFLP pattern analysis and/or combinations of genotyping methods to determine the proportions of new tuberculosis cases that are the result to recently-acquired infection with M. tuberculosis. 2) To define the characteristics of persons who develop tuberculosis from recent infection. 3) To use molecular epidemiologic data to suggest and evaluate specific control interventions. 4) To use local data to construct mathematical models to estimate the effects of new control interventions. 5) To use information from the study sites to develop approaches to the evaluation of candidate tuberculosis vaccines. Taken together this group of aims will serve to provide the epidemiologic basis for tuberculosis control interventions that are specifically targeted to particular epidemiologic circumstances or populations at risk. Component 2: 1) To identify two clonal collections of M. tuberculosis. 2) To use a variety of techniques to genotype the clonal strains. 3) To identify the most parsimonious combination of genotyping markers that can be used to define clonality. 4) To compare the epidemiologic links identified by clonal analysis and by IS6110-based RFLP pattern analysis with the results of intensive epidemiologic investigations. These aims will serve to provide a more accurate and precise means of identifying and tracking strains of M. tuberculosis thereby improving the quality of the information provided by molecular epidemiology. The study is prospective and population-based. Study sites will be San Francisco, Alameda, Contra Costa, and San Mateo counties in the Bay Area and six rural central California counties, Sonoma, Solano, Yolo, Stanislaus, San Joaquin, and Merced. The method of operation will be to obtain M. tuberculosis isolated from all new patients in study areas, perform RFLP pattern analysis on these isolates, identify clusters of cases having the same organism, and perform an extensive detailed interview for cases in clusters and a matched control group to identify epidemiologic connections.